MeltFlex — Furniture Placement

This skill places specific products into a room photo: you provide the room plus one or more reference images of the exact furniture/decor, and MeltFlex composes them into the scene with correct perspective, lighting, and shadows. Same MeltFlex REST API, no extra install.

For general restyling without specific products, use /meltflex:design.

Prerequisites

A MeltFlex API key (mf_sk_..., any active subscription) in MELTFLEX_API_KEY. Get one at <https://www.meltflexai.com/settings>. Each generation costs 10 credits (auto-refunded on failure), billed to the user's account.

Endpoint

POST https://www.meltflexai.com/api/v1/generate

Body:

• prompt (string, required) — placement instruction.
• imageUrl or image — the room photo.
• referenceImageUrls (string[]) — up to 10 public URLs of the products to place. (Or referenceImages as base64 data URLs for local files.)
• referenceProducts (object[], optional) — metadata per reference, e.g. [{"name": "Modular Sofa"}], to help placement.

Response: { "success": true, "image": "data:image/png;base64,...", "creditsUsed": 10 }.

Example

import os, base64, requests

resp = requests.post(
    "https://www.meltflexai.com/api/v1/generate",
    headers={"Authorization": f"Bearer {os.environ['MELTFLEX_API_KEY']}"},
    json={
        "imageUrl": "https://your-cdn.com/empty-room.jpg",
        "prompt": "Place these furniture items naturally in this living room, matching their exact look.",
        "referenceImageUrls": [
            "https://your-cdn.com/modular-sofa.jpg",
            "https://your-cdn.com/oak-coffee-table.jpg"
        ],
        "referenceProducts": [{"name": "Modular Sofa"}, {"name": "Oak Coffee Table"}],
    },
    timeout=180,
)
data = resp.json()
if data.get("success"):
    open("furnished.png", "wb").write(base64.b64decode(data["image"].split(",", 1)[1]))

How to work with the user

1. Need a room photo + at least one product image. If the user names a product but gives no image, ask for a photo or URL of it.
2. Use clean product shots (white/neutral background) for best matching.
3. Name the products in referenceProducts so the AI places them correctly.
4. Up to 10 reference items per call. Report the saved path and credits used.
5. Errors behave the same as /meltflex:design (402 credits, 401 key, 429 rate, 5xx auto-refund).

Tips

• Prefer URLs over base64 for both the room and references — faster, avoids the 15 MB limit.
• For multiple items, describe the desired arrangement in the prompt ("sofa against the back wall, coffee table centered").